Development of an Artificial Intelligence-Driven Adaptive Assessment Model for Formative Evaluation in Nigerian Secondary Schools

This study develops an artificial intelligence-driven adaptive assessment model for formative evaluation in Nigerian secondary schools, addressing critical gaps in contextually appropriate, scalable assessment tools. Traditional formative assessment in Nigerian classrooms is severely limited by class sizes, teacher workload, and limited diagnostic feedback tools. Adaptive assessment systems that tailor question difficulty and feedback to individual learner performance profiles offer a scalable solution. The study employs a design-based research methodology across four iterative cycles. Phase one established design principles through systematic review and stakeholder consultation with 50 secondary school teachers and assessment experts from Lagos, Rivers, and Kano States. Phase two produced a prototype AI-driven assessment system aligned with the Nigerian secondary school curriculum. Phases three and four involved iterative piloting with 300 Junior Secondary School Two students, evaluating system accuracy, teacher usability, and diagnostic validity. Item response theory modelling was used to calibrate the adaptive algorithm. Results demonstrate high diagnostic accuracy, significant reduction in teacher marking workload, and positive student response to personalised feedback. Optimal performance occurred in structured curriculum areas with established item banks. The study presents an original, validated adaptive assessment model designed for the Nigerian secondary school context, contributing to both educational assessment theory and educational technology design practice. Keywords: adaptive assessment, artificial intelligence, formative evaluation, Nigerian secondary school, design-based research